Abdel-Rahman E. M., Landmann T., Kyalo R., Ong'amo G., Mwalusepo S., Sulieman S., Le Rü Bruno. (2017). Predicting stem borer density in maize using RapidEye data and generalized linear models. International Journal of Applied Earth Observation and Geoinformation, 57, p. 61-74. ISSN 0303-2434.
Titre du document
Predicting stem borer density in maize using RapidEye data and generalized linear models
Abdel-Rahman E. M., Landmann T., Kyalo R., Ong'amo G., Mwalusepo S., Sulieman S., Le Rü Bruno
Source
International Journal of Applied Earth Observation and Geoinformation, 2017,
57, p. 61-74 ISSN 0303-2434
Average maize yield in eastern Africa is 2.03 t ha(-1) as compared to gibbal average of 6.06 t ha(-1) due to biotic and abiotic constraints. Amongst the biotic production constraints in Africa, stem borers are the most injurious. In eastern Africa, maize yield losses due to stem borers are currently estimated between 12% and 21% of the total production. The objective of the present study was to explore the possibility of RapidEye spectral data to assess stem borer larva densities in maize fields in two study sites in Kenya. RapidEye images were acquired for the Bomet (western Kenya) test site on the 9th of December 2014 and on 27th of January 2015, and for Machakos (eastern Kenya) a RapidEye image was acquired on the 3rd of January 2015. Five RapidEye spectral bands as well as 30 specttal vegetation indices (SVIs) were utilized to predict per field maize stem borer larva densities using generalized linear models (GLMs), assuming Poisson ('Po') and negative binomial ('NB') distributions. Root mean square error (RMSE) and ratio prediction to deviation (RPD) statistics were used to assess the Models performance using a leave one -out cross-validation approach. The Zero-inflated NB ('ZINB') models outperformed the 'NB' models and stem borer larva densities could only be predicted during the mid growing season in December and early January in both study sites, respectively (RMSE=0.69-1.06 and RPD = 8.25-19.57). Overall, all models performed similar when all the 30 SVIs (non-nested) and only the significant (nested) SVIs were used. The models developed could improve decision making regarding controlling maize stem borers within integrated pest management (IPM) interventions.
Plan de classement
Sciences fondamentales / Techniques d'analyse et de recherche [020]
;
Sciences du monde végétal [076]
;
Télédétection [126]
